#benchmarks

  1. AI models find the errors human peer reviewers miss

    Human peer reviewers found 27% of the major errors planted in three trial reports.

  2. Claude Opus 5.5

    Anthropic's first Claude 5.5 model (Sep 22, 2026).

  3. Coding agents: cost vs speed benchmarks (Stephen Reid)
  4. GPT-6 Astra: A New Generation of Intelligence

    OpenAI model release saturating several major benchmarks (ARC-AGI-3, FrontierMath Tier 4, ExploitBench) in one shot, while also being the…

  5. Introducing GPT-5.5
  6. MOC: Frontier Model Releases

    This map covers major frontier model releases: when each shipped, where it is logged here, and how the frontier has moved on benchmarks…

  7. RoboHarm

    RoboHarm tests whether frontier robot policies refuse physically harmful instructions—and finds that they usually do not.